Ensure Client Liveness: Automatic Failover on Server Unavailability #169
Description
Description:
The d-engine-client must maintain liveness even when server nodes fail or disconnect briefly.
The client should never deadlock or block indefinitely waiting for a server to come back online.
Acceptance Criteria
- Detect when a server endpoint is unavailable (e.g., crash, network partition, overload)
- Automatically fail over to a healthy server node without blocking the client
- Provide clear error messages or status updates for failed nodes
- Ensure client operations continue normally on available nodes
- Include automated tests simulating node failure or disconnect